Generative Adversarial Perturbations

1 code implementation • CVPR 2018 • Omid Poursaeed, Isay Katsman, Bicheng Gao, Serge Belongie

In this paper, we propose novel generative models for creating adversarial examples, slightly perturbed images resembling natural images but maliciously crafted to fool pre-trained models.

General Classification Semantic Segmentation